Unit 6: Funding, staffing, and organizing AI product teams

Learning objectives

Structure product teams for AI work. Navigate the unique staffing needs of AI-augmented teams. Design funding models that support outcome-driven AI product development.

Video script

Reading material

Team composition for AI products

The full AI product team. PM, designer, 2-3 software engineers, 1-2 ML engineers, data analyst. This team can run the full discovery-build-measure cycle for AI features independently. Appropriate for: companies where AI is a core product differentiator.

The augmented product team. PM, designer, 2-3 software engineers, plus access to a shared ML team. The shared ML team handles model development while the product team handles everything else. Appropriate for: companies adding AI features to existing products.

The exploratory team. PM, 1 software engineer, 1 ML engineer. A small, fast-moving team focused on validating AI opportunities through prototypes and experiments. Their output is learning, not production features. Appropriate for: companies early in their AI journey.

The PM's role in AI team dynamics

When ML engineers join the product trio, the collaboration dynamics shift. The PM needs to understand ML development well enough to facilitate productive conversations between ML engineers and the rest of the team.

Key PM responsibilities in an AI team: defining quality thresholds (what "good enough" means for the use case), managing the tension between model quality and shipping speed, ensuring that ML work stays connected to user problems (not just technical optimization), and translating between ML and business stakeholders.

Common pitfall: the PM becomes a translator between ML and the rest of the organization, spending all their time in coordination rather than discovery. If this happens, it means the ML engineer isn't integrated enough into the team. Push for direct ML-to-stakeholder communication with the PM facilitating, not intermediating.

Practical exercise

Exercise: Team design

For an AI product initiative you're considering, design the team.

1. What team composition does this initiative need? (Full, augmented, or exploratory)
2. What specific roles and skills are required?
3. How would you fund this team? (Project-based, team-based, or portfolio approach)
4. What is the team's outcome-based mission?
5. How would the PM and ML engineer collaborate on a typical feature?

Write this up as a one-page team charter.

Leadership reflection: Does your current team structure support or hinder AI product work? What would need to change?